Models Complete Zen LM model family -- 49 models across 10 modalities
Zen LM by Hanzo AI is a comprehensive model family of 49 models across 10 modalities. Each model is built using the Zen MoDE (Mixture of Distilled Experts) methodology — curating the best open-source foundations and fusing them into a unified, high-performance family. 15 models are available via the cloud API at api.hanzo.ai. All open-weight models are on HuggingFace .
The latest generation. Flagship, reasoning, and code models.
Model Parameters Architecture Context Tier Input $/1M Output $/1M zen4-max 1.04T (32B active) MoE 256K ultra max $3.60 $3.60 zen4 744B (40B active) MoE 202K ultra max $3.00 $9.60 zen4-ultra 744B (40B active) MoE + CoT 202K ultra max $3.00 $9.60 zen4-pro 80B (3B active) MoE 131K ultra $2.70 $2.70 zen4-thinking 80B (3B active) MoE + CoT 131K pro max $2.70 $2.70 zen4-mini 8B Dense 40K pro $0.60 $0.60 zen4-coder 480B (35B active) MoE 262K ultra $3.60 $3.60 zen4-coder-pro 480B Dense BF16 262K ultra max $4.50 $4.50 zen4-coder-flash 30B (3B active) MoE 262K pro max $1.50 $1.50
Multimodal, vision, safety, and embedding models.
Model Parameters Architecture Context Tier Input $/1M Output $/1M zen3-omni ~200B Dense Multimodal 202K pro max $1.80 $6.60 zen3-vl 30B (3B active) MoE Vision-Language 131K pro max $0.45 $1.80 zen3-nano 4B Dense 40K pro $0.30 $0.30 zen3-guard 4B Dense 40K pro $0.30 $0.30 zen3-embedding N/A Embedding 8K pro max $0.39 $0.39
Local-first models available on HuggingFace .
Model Parameters Context Description zen-nano 0.6B 32K Ultra-lightweight edge model zen-eco 4B 32K Efficient general-purpose zen 8–32B 32K Standard foundation zen-pro 32B 32K Professional grade zen-max 1.04T MoE 256K Open weights (= zen4-max)
Model Description Status zen-3d Text/image-to-3D generation Coming soon zen-world World simulation Coming soon
Model Parameters Description zen-agent 32B Agentic AI with tool use and planning (preview)
All API models are accessed via the Hanzo SDK. API key prefix: hk-.
from hanzoai import Hanzo client = Hanzo( api_key = "hk-your-api-key" ) response = client.chat.completions.create( model = "zen4" , messages = [{ "role" : "user" , "content" : "Hello!" }], ) print (response.choices[ 0 ].message.content) All locally-runnable models are available on HuggingFace in multiple formats:
Format Use Case Platform SafeTensors Full precision, transformers All GGUF Quantized, llama.cpp/Ollama All MLX Apple Silicon optimized macOS ONNX Cross-platform inference All